Inkling vs Claude Sonnet 4.5 vs Qwen3.6 Max Preview
Inkling comes out ahead, 59 to 55 and 54 on our weighted score, though Qwen3.6 Max Preview is 43% cheaper per token.
- Our pick
Thinking MachinesInkling
59/100- ECI148.6
- Price$3.74 / $9.36
- Context1.05M
AnthropicClaude Sonnet 4.5
55/100- ECI146.8
- Price$3.00 / $15.00
- Context200K
Alibaba (Qwen)Qwen3.6 Max Preview
54/100- ECI149.2
- Price$1.30 / $7.80
- Context262K
Inkling is our pick
Inkling is the better all-round choice, scoring 59/100 against Claude Sonnet 4.5 (55) and Qwen3.6 Max Preview (54). It leads on context window. Claude Sonnet 4.5 wins on inputs & features. Qwen3.6 Max Preview wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityQwen3.6 Max PreviewCapabilities Index (ECI): Qwen3.6 Max Preview 149.2 · Inkling 148.6 · Claude Sonnet 4.5 146.8
- Lowest priceQwen3.6 Max PreviewQwen3.6 Max Preview $2.92 · Inkling $5.14 · Claude Sonnet 4.5 $6.00 per 1M tokens (3:1 blend)
- Longest contextInklingInkling 1,048,576 · Qwen3.6 Max Preview 262,144 · Claude Sonnet 4.5 200,000 tokens
- Widest inputsInkling and Claude Sonnet 4.5Inkling: Text, Images, Audio · Claude Sonnet 4.5: Text, Images, PDFs · Qwen3.6 Max Preview: Text
- Self-hostingInklingPublishes downloadable weights (Apache-2.0)
| Measure | Weight | Inkling | Claude Sonnet 4.5 | Qwen3.6 Max Preview |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 76 | 74 | 77 |
| Price | 25% | 16 | 13 | 28 |
| Inputs & features | 15% | 70 | 80 | 35 |
| Context window | 10% | 61 | 32 | 37 |
| Overall | 100% | 59/100 | 55/100 | 54/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | InklingThinking Machines | Claude Sonnet 4.5Anthropic | Qwen3.6 Max PreviewAlibaba (Qwen) |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 148.6 | 146.8 | 149.2 (best) |
| ECI rank | #57 of 148 | #64 of 148 | #54 of 148 (best) |
| GPQA DiamondGraduate-level science questions | 88.3% (best) | 82.3% | 87.4% |
| FrontierMath Tiers 1–3Research-level mathematics | 33.3% (best) | 23.9% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 88.9% | 77.8% | 91.1% (best) |
| SWE-bench VerifiedFixing real GitHub issues | — | 71.3% | 76.7% (best) |
| SimpleQA VerifiedShort factual questions | 40.3% | 30.7% | 52.0% (best) |
| Price per million tokens | |||
| Input | $3.74 | $3.00 | $1.30 (best) |
| Output | $9.36 | $15.00 | $7.80 (best) |
| Cached input | $0.748 | $0.30 | $0.13 (best) |
| Blended (3:1) | $5.14 | $6.00 | $2.92 (best) |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Official Thinking Machines API | Official Anthropic API | Official Alibaba API |
| Limits | |||
| Context window | 1,048,576 tokens (best) | 200,000 tokens | 262,144 tokens |
| Max output | 1,048,576 tokens (best) | 64,000 tokens | 65,536 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | No |
| PDFs | No | Yes | No |
| Audio | Yes | No | No |
| Video | No | No | No |
| Reasoning | Yeslow · medium · high · xhigh · max | Yes | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | OpenApache-2.0 | Proprietary | Proprietary |
| API model ID | thinkingmachines/Inkling:peft:262144 | claude-sonnet-4-5 | qwen3.6-max-preview |
| API providers | 23 | 28 (best) | 10 |
| Released | Jul 15, 2026 | Sep 29, 2025 | Apr 20, 2026 |
| Knowledge cutoff | — | Jul 31, 2025 | Apr 2025 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Inkling$56.12
- Claude Sonnet 4.5$60.00
- Qwen3.6 Max Preview$28.60
Which should you choose?
Which is better: Inkling, Claude Sonnet 4.5 or Qwen3.6 Max Preview?
Inkling is the better all-round choice, scoring 59/100 against Claude Sonnet 4.5 (55) and Qwen3.6 Max Preview (54). It leads on context window. Claude Sonnet 4.5 wins on inputs & features. Qwen3.6 Max Preview wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, Inkling, Claude Sonnet 4.5 or Qwen3.6 Max Preview?
Qwen3.6 Max Preview is cheaper at $1.30 input / $7.80 output per million tokens (official Alibaba API price). Inkling costs $3.74 input / $9.36 output per million tokens (official Thinking Machines API price); Claude Sonnet 4.5 costs $3.00 input / $15.00 output per million tokens (official Anthropic API price). At a typical mix of three input tokens to one output token, that is $2.92 per million tokens for Qwen3.6 Max Preview versus $5.14 for Inkling (1.8× as much) and $6.00 for Claude Sonnet 4.5 (2.1× as much).
Which scores higher on benchmarks?
Qwen3.6 Max Preview scores higher on the Capabilities Index (ECI): Qwen3.6 Max Preview 149.2 (#54 of 148), Inkling 148.6 (#57 of 148) and Claude Sonnet 4.5 146.8 (#64 of 148). The confidence ranges of the top two overlap (147.6–152.0 vs 145.7–150.6), so treat the gap as small. On individual benchmarks: GPQA Diamond — Inkling 88.3%, Qwen3.6 Max Preview 87.4%, Claude Sonnet 4.5 82.3%; OTIS Mock AIME 2024–2025 — Qwen3.6 Max Preview 91.1%, Inkling 88.9%, Claude Sonnet 4.5 77.8%; SimpleQA Verified — Qwen3.6 Max Preview 52.0%, Inkling 40.3%, Claude Sonnet 4.5 30.7%.
Which is better for coding?
There are no published SWE-bench Verified results for Inkling yet, so there is no like-for-like coding score. On overall capability, Qwen3.6 Max Preview leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
Inkling has the largest context window at 1,048,576 tokens, against 262,144 for Qwen3.6 Max Preview and 200,000 for Claude Sonnet 4.5. Maximum output per response: Inkling up to 1,048,576, Claude Sonnet 4.5 up to 64,000, Qwen3.6 Max Preview up to 65,536 tokens.
Which can read images, PDFs, audio or video?
Inkling accepts text, images and audio; Claude Sonnet 4.5 accepts text, images and PDFs; Qwen3.6 Max Preview accepts text. Inkling handles the widest range of inputs.
Are any of these open source?
Inkling publishes its weights (Apache-2.0) and can be self-hosted; Claude Sonnet 4.5 and Qwen3.6 Max Preview is proprietary.
Which is newer?
Inkling is the newest, released Jul 15, 2026. Qwen3.6 Max Preview came out Apr 20, 2026; Claude Sonnet 4.5 came out Sep 29, 2025. Knowledge cutoff: Claude Sonnet 4.5 Jul 31, 2025, Qwen3.6 Max Preview Apr 2025.
How do you decide the winner?
Each model gets a 0–100 score on capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Oct 4, 2026.
